April 2026 · SLA Theory · Methodology
This article explains the methodological spectrum that the Edooqoo Learning Pacing slider exposes inside the Dynamic Student Learning Model (DSLM). The slider does not change the worksheet engine — it only changes how the AI plans phases and next steps for an individual adult learner. Below is the underlying second-language acquisition (SLA) research that informs both ends of that spectrum.
"Pacing" in this context is not lesson speed. It is the ordering principle behind which language items are introduced first, when productive use is required, and how tightly content is bound to the learner's professional or personal domain. Two well-documented poles dominate the literature:
Most adult learners benefit from a balance, but the optimal blend depends on level, deadline, and goal. The slider lets the teacher set this blend explicitly per student.
Adult ESL is not a homogeneous setting. A B1 software engineer with a meeting in 14 days has different needs than an A2 retiree learning English for travel six months out. Methodological monism — applying one methodology to every learner — has been criticized in the SLA literature for at least two decades2. The Edooqoo pacing slider operationalizes that critique inside the planner.
Dulay and Burt (1974) found that L2 learners acquire certain grammatical morphemes in a predictable order largely independent of L11. Krashen incorporated this into his Monitor Model3. Goldschneider and DeKeyser (2005) ran a meta-analysis confirming that this morpheme order is statistically stable across studies and largely explained by salience, frequency, regularity, and semantic complexity4. The practical implication: pushing A1/A2 learners into productive use of passives, reported speech, or mixed conditionals tends to fail because those structures sit late in the developmental order.
Krashen's Input Hypothesis (1985) argues that acquisition is driven by exposure to input slightly above the learner's current level (i+1)5. The hypothesis is not uncontested — Lichtman and VanPatten (2021) reviewed forty years of evidence and concluded that comprehensible input is necessary but not sufficient on its own6. The scientific pole nonetheless treats sufficient, level-appropriate input as a precondition before output is required.
Sweller's Cognitive Load Theory and its applications to second-language instruction (Roussel, Sweller and Tricot, 2022; Bledsoe and Richardson, 2020) emphasize that working memory is finite and that overloading it with simultaneous new vocabulary, new grammar, and unfamiliar topics blocks learning78. Edooqoo enforces this by limiting the number of new grammar points per generated step and by anchoring all examples in a single domain familiar to the learner.
Swain (1985) argued that producing language — not just receiving it — drives certain types of acquisition, especially syntactic processing9. Subsequent work, including Peker (2020), supports the view that pushed output, when scaffolded, accelerates accuracy10. The pragmatic pole front-loads output for adults whose context demands it.
Ellis (2009, 2014, 2017) defined and refined Task-Based Language Teaching (TBLT), in which lessons are organized around real communicative tasks rather than around grammar points21112. Under TBLT, a lesson titled "Explaining a long-standing bug to your CTO at the standup" is preferred over "Present perfect continuous practice" because the task — not the form — is the unit of design.
Hutchinson and Waters (1987) established ESP as needs-driven instruction tied to the learner's domain13. Basturkmen (2022) and Yan (2025) summarize the contemporary evidence base and confirm that ESP improves engagement and retention in adult professional learners1415. The pragmatic pole leans heavily on ESP framing from lesson one.
Most adults learn fastest with a balanced mix: enough input scaffolding to respect developmental orders, but immediate task relevance so the language is usable in the learner's life within days, not months. The Edooqoo "Balanced" preset (31–69 on the slider) corresponds to this middle ground: natural order is respected, but every lesson is anchored in the learner's profession, industry, or stated personal goal.
Cepeda, Pashler, Vul, Wixted and Rohrer (2006) ran a meta-analysis showing large effects of spaced over massed practice16. Kim and Webb (2022) confirmed substantial effects in second-language vocabulary acquisition specifically17. The DSLM planner returns to weak skills after one to three intervening lessons, not in the next lesson.
Nakata and Suzuki (2019) and Suzuki, Nakata and DeKeyser (2019) demonstrated that interleaving exercise types within a session improves retention compared to blocked practice1819. Edooqoo lessons therefore mix vocabulary-focused and grammar-focused exercises rather than batching them.
Roediger and Karpicke (2006) and Karpicke and Roediger (2008) established that retrieval — being asked to produce — is more effective for long-term retention than re-exposure2021. Each generated lesson includes at least two productive exercises (answer-questions, dialogue, discussion, or open fill-in-blanks).
Knowles (1980) framed adult learning as needs-driven, autonomous, and experience-based22. Clardy (2005) added critical nuance about its limits23. Edooqoo lesson titles, examples, and goals follow andragogical conventions: adult professional tone, immediate relevance, no childlike imagery.
The slider value is read by two edge functions: generate-curriculum-phases and generate-timeline. It changes the prompt sent to the planner — it does not change the worksheet generation engine itself. Specifically:
The slider does not affect billing, token consumption, or worksheet rendering. It is a planning-layer control only.
Beyond the manual slider, Edooqoo continuously recalculates a recommended pacing value from signals. When the recommendation diverges from the current value by more than 15 points, you receive a pacing proposal with a written rationale. You always decide whether to accept it.
Time remaining to the goal deadline sets a minimum pacing the system will not go below:
| Days remaining | Minimum pacing |
|---|---|
| ≤ 14 | 100 (forced Pragmatic) |
| ≤ 30 | ≥ 80 |
| ≤ 60 | ≥ 60 |
| ≤ 90 | ≥ 40 |
| > 90 | signals-driven |
| Signal (from Self-Profile + events) | Effect |
|---|---|
| Anxiety reported high in learning_obstacles | −10 |
| Perfectionism reported high | −10 |
| motivation_driver = career or exam | +10 |
| time_availability_per_week < 3 h | −15 |
| time_availability_per_week ≥ 7 h | +10 |
| completion_velocity (events) > 1.2 | +15 |
| completion_velocity < 0.7 | −15 |
The final value is clamped to the 0–100 range and never falls below the Step 1 floor when the deadline is ≤ 30 days.
If the new value differs from the current one by more than 15 points, the system creates a pacing proposal visible in the teacher dashboard with a written rationale (e.g., "deadline pressure rising + low time availability — moving toward TBLT"). You accept, reject, or ignore. Nothing changes automatically without your decision.